OMG Meeting

This trip report covers the high points of the OMG Seattle meeting and their relevance to DARPA.  We view OMG meetings as a great way for DARPA and industry to transfer technology and expertise.  There are numerous other connection points beyond the few reported here where DARPA and OMG could create synergy, e.g., in areas like security, simulation, GIS, QoS, real-time, etc.  A small team of people just barely begins to cover the possibilities.

• Internet SIG (minutes: http://www.objs.com/isig/mtg19.htm; co-chairs Craig Thompson OBJS and Shel Sutton, MITRE):
• HTTP-NG presentation.  Bill Janssen (Xerox PARC) described the W3C project to slide an ORB under the IETF HTTP protocol, aiming to eventually make this an IETF standard.  He is looking for the rendezvous point in OMG and asked Thompson to chair a Task Force to make this happen; no commitments made.  This *is* an important project aimed at wedding the web and ORBs; DARPA should track this project.
• Web Object Model presentations.  Shel Sutton (MITRE) described a host of XML based proposed specs that W3C is looking at.  Frank Manola (OBJS) made further sense of these by showing how they might be viewed as a decentralized object model.
• Web/OMG Integration Architectures Working Group. Frank Manola (OBJS) led this working group, whose goal is to develop a coherent strategy for integrating Web and OMG technologies (rather than allowing this to happen piecemeal). The importance of this is illustrated by the accelerating number of OMG technologies which involve integration with Web technologies, including:
• HTTP-NG (see above)
• the use of XML in XMI (OMG metadata exchange submission), Tagged Data Facility, Common (Data) Warehouse Metadata RFP, and CORBA Components submission
• the incorporation of Web authoring requirements in the OMG Document Management RFP
• the requirement for OMG's revised Interoperable Naming Service to handle URL to object reference mappings
Tracking all this required attending various OMG meetings:  ORBOS, OA&D, Document Management.  DARPA should be aware of the general interest in OMG in web-object interoperability and in some specific specifications like the joint components submission.
• CSCW Working Group (minutes:  http://www.objs.com/isig/wg-cscw01.html).  Henry Rothkopf (MITRE, DARPA OMG representative) led a working group on collaboration.  The idea is to develop a framework for slotting collaboration technologies into the OMG OMA (his presentations are here and here).  Henry is setting up a good connection for DARPA JFACC collaboration services, which should brief OMG at one of the next OMG meetings.  IC&V could take advantage of this too.
• Agents Working Group (minutes: http://www.objs.com/isig/wg-agents01.html).  This new working group (which Thompson co-chairs) made a lot of progress for a first meeting.  FIPA and OMG MASIF presentations indicate these specs are complementary. Thompson provided a presentation on a Strawman Agents Reference Architecture, introducing the idea of the coABS grid to OMG; Jeff Bradshaw (Boeing) presented other coABS work; and Roger Burkhart and James Odell presented on emergent behavior and swarms.  Doyle Weishar attended as did Sun Chelmsford researchers.
• Object and Reference Model Subcommittee: The work on a Next Generation OMA Guide is in limbo with its chair person Kevin Tyson's recent resignation right before the meeting.  We could easily pick up the work and gear this toward rendezvous with DARPA ISO architecture directions.  Meanwhile, Frank Manola participated in the meeting of the Object Model Working Group, tasked to define a new OMG Core Object Model, which plays a key role in the specification of the Next Generation Object Management Architecture. The group discussed a proposed draft (including text prepared by Frank), and its alignment with the current OMG IDL object model and extensions such as Objects-By-Value and multiple interfaces, with a goal of having a fairly complete specification by the January meeting.
• Business Object Domain Task Force (BODTF): A new working group was created called the "Business Object Initiative Working Group" to build consensus on proceeding with the "business objects/components" idea, something not present in the earlier, now defunct,  BOCA (Business Objects Component Architecture) work. The group presented three new draft RFPs all on aspects of UML for business objects to the Business Object and Analysis & Design Task Forces.   Separately, a "Domain Value Types" group was established, with the objective to define common value types like units of measure.  Agreement in this area would reduce semantic interoperability problems that plague us all.  Frank Manola contributed to get this group off to a good start.
• past and future OMG trip reports are/will be at http://www.objs.com/omg/omg.html